More than once it has been stated that autism is a condition which is defined exclusively by the overt presentation of a specific set of behaviours anchored to a specific period of onset. The description of those behaviours might be due for a bit of an overhaul in the coming years with the various proposals attached to schedules like DSM-V, but essentially the same diagnostic features remain. When trying to assess core behaviours, there are a variety of schedules 'on the market'. Some of the gold standards like ADI-R and ADOS have appeared fairly frequently on this blog. I say 'gold standard' because when it comes to providing some standardised, objective (almost) data based on the defining behavioural and developmental features, these two instruments in particular have stepped up to the mark. Notable also because the level of autistic behaviours according to domain can be scored on each instrument and compared against cut-off values. There are other instruments that fit into the 'gold standard' category but I will stick with these two for now.
Fine for assessing behaviour in autism, but what about assessing change in behaviour in autism?
'Change' in symptoms in autism occurs quite readily for a variety of reasons. Aside from the early diagnostic instability which I have discussed in a previous post, change can occur as a result of maturation (remember my puberty post), as result of social and environmental factors and has also been suggested as a result of intervention (in some cases). I would imagine therefore that most people would not view autism as a static entity but something rather more dynamic.
The problems however start when change is evaluated; bearing in mind that change in autism normally is not just confined to the 'core' diagnostic symptoms but often, other peripheral presentations (indeed, one might argue that it is the knock-on effect from peripheral changes to core symptoms which is important). What do we do then?
The appropriateness of outcome measure has been discussed in the scientific literature. This paper discussed various issues including 'outcome measures sensitive to individual change'. The number of times I have come across papers that have found no significant effects for one thing or another but have then commented that the outcome measure may have missed target behaviours is beyond a joke. The problem is that when it comes to assessing change there is no universal 'gold standard' instrument. OK some researchers (including ourselves) have used instruments like ADOS. On the whole these are OK as long as you give a long enough period of time between assessment sessions so as not to encounter any practice effects. Bearing in mind also that ADOS was never really designed with measuring change in mind, so we are perhaps trying to fit a square block into a round hole.
The instruments that have been configured with change in mind, whilst also OK, also have their short-comings. The ATEC (Autism Treatment Evaluation Checklist) has been around for quite a while and used in quite a few studies, mostly concerned with the more 'biomedical' interventions. I have to say that I was always a little hesitant with the ATEC, thinking that perhaps it was not as standardised as it should be. Like a snob looking down at a fish and chip supper. My opinion has however changed in recent years and especially following the thumbs up from this study a few weeks back. Don't get me wrong, there are still quite a few shortcomings, particularly associated with the use of a 3-point Likert scale/item to assess behaviour and also in the definition of some of the items (how do you define 'constipation' and what if constipation to you is not the same as constipation to me?)
Another instrument, which I will readily admit has been a favourite of mine, is the Behaviour Summarized Evaluation (BSE) schedule. For some reason nearly all the links on the web mis-spell this instrument as the behavioural summarized, etc. I know, I know BSE... for us here in the UK conjours up cows in a pretty dire state and lots of quite distressing bonfires - not one of our finest hours. But the BSE instrument for autism is a pretty good scale and has the evidence to back it up (here and here). I again admit that for one of our first papers on the use of a gluten-free diet for autism, we did use the BSE and I have to say it performed pretty well. If I was to criticise it, I would have to say that it did not include some important features like gastrointestinal / functional bowel problems, which we were keen also to look at in that particular study.
From what I understand there are other schedules in development. My Christmas wish-list of the things I would like to see in any schedule to assess change are:
- A 5- or 7-point Likert scale/item which, like the ADOS and ADI-R offers specific descriptions for each option.
- A schedule which includes both core and peripheral symptoms (e.g. inattention, hyperactivity, impulsivity).
- A schedule which also includes items relating to more somatic features (e.g. functional bowel problems, sleeping problems, motor coordination and gait problems).
- A schedule which again like ADOS and ADI-R can be standardised across raters.
The weird thing is that all these elements are already there across the various instruments but not in one single instrument - maybe a revamp of the ADOS or ADI-R is in order? Assuming such a schedule ever appears, and some consensus on use is made (making it open-access perhaps?) it would then provide a platform for evaluating both change as a function of the natural progressions of autisms and change as a function of the various interventions.
Take it away Dave - Genius.